Abstract:In the era of Industry 4.0, Additive Manufacturing (AM), particularly metal AM, has emerged as a significant contributor due to its innovative and cost-effective approach to fabricate highly intricate geometries. Despite its potential, this industry still lacks real-time capable process monitoring algorithms. Recent advancements in this field suggest that Melt Pool (MP) signatures during the fabrication process contain crucial information about process dynamics and quality. To obtain this information, various sensory approaches, such as high-speed cameras-based vision modules are employed for online fabrication monitoring. However, many conventional in-depth analyses still cannot process all the recorded data simultaneously. Although conventional Image Processing (ImP) solutions provide a targeted tunable approach, they pose a trade-off between convergence certainty and convergence speed. As a result, conventional methods are not suitable for a dynamically changing application like MP monitoring. Therefore, this article proposes the implementation of a Tunable Deep Image Processing (TDIP) method to address the data-rich monitoring needs in real-time. The proposed model is first trained to replicate an ImP algorithm with tunable features and methodology. The TDIP model is then further improved to account for MP geometries and fabrication quality based on the vision input and process parameters. The TDIP model achieved over 94% estimation accuracy with more than 96% R2 score for quality, geometry, and MP signature estimation and isolation. The TDIP model can process 500 images per second, while conventional methods taking a few minutes per image. This significant processing time reduction enables the integration of vision-based monitoring in real-time for processes and quality estimation.
What problem does this paper attempt to address?
This paper attempts to solve the problem of real - time melt pool (MP) monitoring in the metal additive manufacturing (AM) process. Specifically, the existing melt pool monitoring methods have problems of slow speed and insufficient accuracy when dealing with a large amount of data, and cannot meet the requirements for real - time monitoring in the Industry 4.0 era. To overcome these problems, the paper proposes a Tunable Deep Image Processing (TDIP) method to achieve efficient real - time melt pool monitoring and quality assessment.
### Main problems in the paper:
1. **Requirement for real - time melt pool monitoring**: In the metal additive manufacturing process, the dynamic changes of the melt pool contain a large amount of process information. Real - time monitoring of this information is crucial for ensuring manufacturing quality and optimizing process parameters.
2. **Limitations of existing methods**:
- Although traditional Image Processing (ImP) methods can provide targeted adjustments, there is a trade - off between convergence speed and accuracy, and it is difficult to adapt to dynamically changing application scenarios.
- Although existing machine - learning methods can improve the processing speed, they lack flexibility and cannot be adjusted according to user needs.
### Solutions:
The paper proposes the TDIP model, aiming to achieve efficient and flexible real - time melt pool monitoring through deep - learning technology. The main features of the TDIP model include:
- **High - speed processing ability**: The TDIP model can process 500 images per second, far exceeding the processing speed of traditional methods.
- **Tunability**: By introducing hyper - parameters (such as F_size, M_ratio, T_ratio, etc.), the model can be adjusted according to user needs to adapt to different application scenarios.
- **Multi - task processing**: The TDIP model can not only isolate the melt pool features, but also estimate the melt pool geometric parameters and the final printing quality (height, width, depth, etc.).
### Experimental results:
- **TDIP V.1**: In the image - processing task, TDIP V.1 has reached an R² score of 97%, and the processing speed is about 4,800 times faster than that of traditional methods.
- **TDIP V.2**: With the addition of the melt pool geometric - parameter - estimation function, the processing speed has been increased by about 120,000 times, and the accuracy rate of melt pool geometric - parameter estimation is as high as 99.5%.
- **TDIP V.3**: The function of the model has been further expanded, and it can simultaneously process image denoising, melt pool geometric - parameter estimation and final printing - quality assessment, significantly improving the ability of real - time monitoring and quality control.
Through these improvements, the TDIP model provides an efficient, flexible and accurate solution for real - time melt pool monitoring in the metal additive - manufacturing process.